Real-time personal healthcare data analysis using edge computing for multimodal wearable sensors
Guren Matsumura, Satoko Honda, Takamasa Kikuchi, Yuuki Mizuno, Hyuga Hara, Yoshiki Kondo, Haruki Nakamura, Shin Watanabe, Kiyoshi Hayakawa, Kohei Nakajima, Kuniharu Takei
Abstract
Wireless, multimodal, wearable sensor patches can perform remote diagnosis and monitoring. By integrating multiple sensors, multimodal, wearable patches generate large datasets. Complex data correlation and analyses of large datasets for real-time, automatic edge-type systems are challenging. Here, we present an integrated wearable sensor patch with edge computing for remote healthcare applications powered by reservoir computing. This sensor patch system is integrated with flexible sensors for electrocardiography, respiration, skin temperature, and skin humidity. Vital and activity data can be applied to monitor coughing, arrhythmias, and posture using a machine learning algorithm. Importantly, all measurements, wireless data transmission, and real-time data analyses are processed in a smartphone as an edge-computing system, which allows quick, real-time feedback to users without using a cloud network system.